Cybersecurity experts report a significant rise in LLMjacking, a criminal trend where attackers steal AI credentials and API keys to illicitly access expensive computing resources and models.
Key Points
- LLMjacking involves unauthorized parties hijacking AI accounts to perform computing tasks or run malicious models at the victim's expense.
- Google Threat Intelligence Group reports a major increase in these attacks throughout 2026, targeting models from providers like OpenAI, Anthropic, and Google.
- Sysdig researchers estimate that unauthorized AI usage can result in financial losses ranging from $46,000 to over $100,000 per day for enterprise victims.
- Stolen credentials are being sold on underground markets for up to 97% off the standard cost of legitimate AI services.
- Businesses are advised to implement zero-trust frameworks, rotate API keys, and conduct frequent security audits to prevent credential theft and account exploitation.